Provably End-to-end Label-Noise Learning without Anchor Points

Abstract

In label-noise learning, the transition matrix plays a key role in building statistically consistent classifiers. Existing consistent estimators for the transition matrix have been developed by exploiting anchor points. However, the anchor-point assumption is not always satisfied in real scenarios. In this paper, we propose an end-to-end framework for solving label-noise learning without anchor points, in which we simultaneously optimize two objectives: the cross entropy loss between the noisy label and the predicted probability by the neural network, and the volume of the simplex formed by the columns of the transition matrix. Our proposed framework can identify the transition matrix if the clean class-posterior probabilities are sufficiently scattered. This is by far the mildest assumption under which the transition matrix is provably identifiable and the learned classifier is statistically consistent. Experimental results on benchmark datasets demonstrate the effectiveness and robustness of the proposed method.

References in corpus (19)

• Robust Recovery of Subspace Structures by Low-Rank Representation
• Learning to Reweight Examples for Robust Deep Learning
• Classification with Noisy Labels by Importance Reweighting
• Training Deep Neural Networks on Noisy Labels with Bootstrapping
• DivideMix: Learning with Noisy Labels as Semi-supervised Learning
• A Closer Look at Memorization in Deep Networks
• WebVision Database: Visual Learning and Understanding from Web Data
• Dimensionality-Driven Learning with Noisy Labels
• Are Anchor Points Really Indispensable in Label-Noise Learning?
• Masking: A New Perspective of Noisy Supervision
• Part-dependent Label Noise: Towards Instance-dependent Label Noise
• Robust Volume Minimization-Based Matrix Factorization for Remote Sensing and Document Clustering
• A Survey of Label-noise Representation Learning: Past, Present and Future
• On Identifiability of Nonnegative Matrix Factorization
• Robustness of Conditional GANs to Noisy Labels
• SIGUA: Forgetting May Make Learning with Noisy Labels More Robust
• Learning with Confident Examples: Rank Pruning for Robust Classification with Noisy Labels
• Unbiased Risk Estimators Can Mislead: A Case Study of Learning with Complementary Labels
• Generalization Bounds for Neural Networks via Approximate Description Length

Cited by in corpus (6)

• Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations
• Sample Selection with Uncertainty of Losses for Learning with Noisy Labels
• Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization
• Clusterability as an Alternative to Anchor Points When Learning with Noisy Labels
• A Second-Order Approach to Learning with Instance-Dependent Label Noise
• Policy Learning Using Weak Supervision